Roman Sakhno - Fotolia

Why enterprises choose multi-cloud environments

Multi-cloud environments have become more popular with enterprises, but this isn't a vendor lock-in issue. Find out how and why enterprises adopt multi-cloud computing.

Enterprises can use multiple clouds to construct a diverse and complex cloud stack, and it's not a new idea. Once the technology caught up to the needs of enterprises, multi-cloud computing became an appealing option for IT professionals looking to get the most out of their IT architecture.

This type of resource pooling is a popular choice among enterprises because of the benefits of multi-cloud computing, including the ability to pick and choose between cloud services based on need and fit. To get the most out of a multi-cloud environment, enterprises should dive deeper into the components that make up their cloud architecture, as well as the reasons so many organizations have implemented multi-cloud.  

Daryl Plummer, VP and distinguished analyst at Gartner, says that in order for users to understand multi-cloud, they must not confuse it with hybrid and distributed cloud architectures. Here's how he breaks it down:

  • Hybrid cloud crosses a barrier between cloud styles, e.g., private and public.
  • Multi-cloud goes across two or more public cloud providers.
  • Distributed cloud brings public cloud services to you wherever you want them.

Once you can distinguish what is or is not a multi-cloud environment, you'll be better equipped to build it.

The layers of multi-cloud

Multi-cloud architectures can be broken down into three notable layers. The first layer deals with multi-cloud administration of SaaS. For example, if an organization uses three SaaS applications, they probably won't all come from the same provider.

Daryl Plummer, distinguished VP and chief Gartner fellowDaryl Plummer

"[That] means you have to administer the bills from multiple cloud providers, you have to do [SLA, data and policy] management," he said. "All of these things are multi-cloud issues forced on you by the fact that SaaS comes from different providers, by and large."

The second layer is related to management and governance tooling. It's most often used when an enterprise needs visibility into the services it runs in different clouds -- from IaaS to SaaS -- even if those clouds don't have anything to do with one another, Plummer said. This scenario has become more common as enterprises increase the number of clouds they use.

It is difficult for enterprises to get everything they need from a single provider. Yet, IT teams still want to centrally manage their services. An organization might prefer AWS for one use case and Google Cloud Platform (GCP) for another but still want to manage VMs in both clouds from one console, Plummer said.

The final, foundational layer involves running workloads that span infrastructure services on different providers' clouds. In these instances, Plummer says users should ask themselves: Are those services working independently from one another or are they working jointly to solve a problem across clouds? If it's the former, an organization might want to avoid this type of multi-cloud setup.

While these situations are still rare, tools and services have emerged that make these situations a more ordinary part of the way enterprises operate, Plummer said. These products include Google Anthos and IBM Cloud Satellite, as well as as-a-service offerings from Red Hat and VMware.

"What those [services] all promote, is the idea if you have a set of skills and solution built in this platform … and it can run on different clouds, then your work should be able to run there too, virtually unchanged," Plummer said.

Why deploy a multi-cloud environment?

A multi-cloud environment may take some work to deploy because of the difference between clouds, but it still has a role in modern cloud computing. Plummer compared it to choosing a restaurant.

If you want a quick meal, where you get the same result every time, you go to a fast food restaurant, such as McDonald's. While you can order whatever you like, you cannot customize it much, beyond adding cheese or condiments. This is similar to the way cloud computing works in its most generic form. However, if you want to go deeper into the cloud and have more control over the results, it would be similar to going to an upscale steakhouse and having your steak cooked to your liking.

It's no secret that all the major IaaS platforms have their strengths and weaknesses, but which cloud is the right one?

"The answer is the same as restaurants -- no one would want to want to eat at one restaurant for the rest of their life," Plummer said. A multi-cloud environment provides choice. You can evaluate different pricing models, quality levels, as well as exposure to different tools and services.

The myth of vendor lock-in

All organizations want to avoid vendor lock-in -- or so you might think.

"It's an illusion," Plummer said. "[Most] enterprises really love being locked-in, because they always want to standardize everything."

The shift to multi-cloud isn't driven by lock-in fears. Instead, it comes back to having the flexibility to choose the right cloud for the right situation, whenever you want. "If I'm equally adept at working on GCP as I am on Alibaba's cloud, then who cares which one I'm working on? I can use either one interchangeably because I know how to do everything [that] I need to do on either one -- if I have the skills."

And while it's not the primary benefit of being multi-cloud, having that diverse skill set can help guard against lock-in fears. It gives an organization a backup option in case the relationship with its current primary cloud provider sours.

"If you have to get out, how fast can you get out? Well, if you're not prepared with skills on another cloud, it's going to take you a quite a while," said Plummer. If your team is already proficient on multiple clouds, your enterprise is already prepared to handle a multi-cloud environment.

Next Steps

4 common traps to avoid in a multi-cloud strategy

Dig Deeper on Cloud deployment and architecture

Data Center